Reduce time for enzyme import

[wsmoses]

Before this change:

julia> @time_imports using Enzyme
┌ Warning: The project dependencies or compat requirements have changed since the manifest was last resolved.
│ It is recommended to `Pkg.resolve()` or consider `Pkg.update()` if necessary.
└ @ Pkg.API /Applications/Julia-1.10.app/Contents/Resources/julia/share/julia/stdlib/v1.10/Pkg/src/API.jl:1807
Precompiling Enzyme
        Info Given Enzyme was explicitly requested, output will be shown live 
WARNING: method definition for PrimalErrorThunk at /Users/wmoses/git/Enzyme.jl/src/compiler.jl:6105 declares type variable World but does not use it.
  2 dependencies successfully precompiled in 112 seconds. 16 already precompiled.
  1 dependency had output during precompilation:
┌ Enzyme
│  [Output was shown above]
└  
      1.3 ms  EnzymeCore
      1.1 ms  CEnum
      6.8 ms  Preferences
      0.3 ms  LazyArtifacts
      0.5 ms  JLLWrappers
               ┌ 3.3 ms LLVMExtra_jll.__init__() 69.32% compilation time
     16.4 ms  LLVMExtra_jll 88.20% compilation time
               ┌ 0.1 ms LLVM.__init__() 
     20.0 ms  LLVM
               ┌ 1.0 ms Enzyme_jll.__init__() 
      2.0 ms  Enzyme_jll
      0.4 ms  ExprTools
               ┌ 0.0 ms TimerOutputs.__init__() 
     10.8 ms  TimerOutputs
      0.4 ms  Scratch
               ┌ 0.4 ms GPUCompiler.__init__() 
     66.1 ms  GPUCompiler 0.14% compilation time
      0.5 ms  Reexport
      0.4 ms  StructIO
               ┌ 0.0 ms ObjectFile.__init__() 
      5.6 ms  ObjectFile
               ┌ 6.4 ms Enzyme.API.__init__() 
               ├ 0.2 ms Enzyme.Compiler.JIT.__init__() 
               ├ 0.0 ms Enzyme.Compiler.FFI.BLASSupport.__init__() 
               ├ 1.9 ms Enzyme.Compiler.FFI.__init__() 
               ├ 0.3 ms Enzyme.Compiler.__init__() 
   1429.7 ms  Enzyme 0.00% compilation time

After this change:

julia> @time_imports using Enzyme
      1.4 ms  EnzymeCore
      1.3 ms  CEnum
      7.9 ms  Preferences
      0.4 ms  LazyArtifacts
      0.4 ms  JLLWrappers
               ┌ 3.3 ms LLVMExtra_jll.__init__() 63.70% compilation time
     17.1 ms  LLVMExtra_jll 87.53% compilation time
               ┌ 0.2 ms LLVM.__init__() 
     19.8 ms  LLVM
               ┌ 1.2 ms Enzyme_jll.__init__() 
      1.9 ms  Enzyme_jll
      0.4 ms  ExprTools
               ┌ 0.0 ms TimerOutputs.__init__() 
     11.2 ms  TimerOutputs
      0.4 ms  Scratch
               ┌ 0.4 ms GPUCompiler.__init__() 
     71.0 ms  GPUCompiler 0.06% compilation time
      0.4 ms  Reexport
      0.4 ms  StructIO
               ┌ 0.0 ms ObjectFile.__init__() 
      5.4 ms  ObjectFile
               ┌ 1.4 ms Enzyme.API.__init__() 
               ├ 0.1 ms Enzyme.Compiler.JIT.__init__() 
               ├ 0.0 ms Enzyme.Compiler.FFI.BLASSupport.__init__() 
               ├ 1.9 ms Enzyme.Compiler.FFI.__init__() 
               ├ 0.0 ms Enzyme.Compiler.__init__() 
     75.3 ms  Enzyme 0.10% compilation time

[github-actions]

Benchmark Results

	main	3c4c2d1...	main/3c4c2d1d85cdb9...
basics/overhead	4.34 ± 0.01 ns	4.34 ± 0.01 ns	1
time_to_load	0.421 ± 0.00089 s	0.455 ± 0.0059 s	0.925

Benchmark Plots

A plot of the benchmark results have been uploaded as an artifact to the workflow run for this PR.
Go to "Actions"->"Benchmark a pull request"->[the most recent run]->"Artifacts" (at the bottom).